我最近开始尝试创建一些与RESTful web接口配合的应用程序，不过我担心每次按F5运行一系列测试时都会对其服务器造成压力。
基本上，我需要获取一系列web响应以便测试是否正确解析了不同的响应，而不是每次都访问他们的服务器。因此，我想可以做一次这样的操作，保存XML文件，然后在本地工作。
但是，我不知道如何“mock”一个WebResponse，因为（据我所知）它们只能由WebRequest.GetResponse实例化。
你们如何处理这种mock情况？你们有这样做吗？我真的不喜欢自己在攻击他们的服务器：S 我不想改变代码太多，但我认为肯定有一种优雅的方法可以解决这个问题...
Will的答案给了我很大的启示，我知道自己忽略了一个根本性的点！
一旦我完成了代码，我就会贴一些样本。
我在寻找实现相同需求时发现了这个问题。 无法在任何地方找到答案，但是经过更深入的挖掘后发现，.Net Framework具有内建支持。
您可以使用WebRequest.RegisterPrefix方法注册一个工厂对象，当使用该前缀（或url）时，WebRequest.Create将调用该对象。 工厂对象必须实现IWebRequestCreate接口，该接口具有单个Create方法，该方法返回一个WebRequest对象。 在此处，您可以返回您的模拟WebRequest对象。
我在这里放置了一些示例代码。
HttpWebResponse来查看状态码...那么你需要查看我下面的答案！ - escape-llc以下是一种不需要模拟的解决方案。您需要实现WebRequest的三个组件：IWebRequestCreate、WebRequest和WebResponse。请参见下面的示例。我的示例生成失败的请求（通过抛出WebException），但应该能够将其调整为发送“真实”的响应：
class WebRequestFailedCreate : IWebRequestCreate {
HttpStatusCode status;
String statusDescription;
public WebRequestFailedCreate(HttpStatusCode hsc, String sd) {
status = hsc;
statusDescription = sd;
}
#region IWebRequestCreate Members
public WebRequest Create(Uri uri) {
return new WebRequestFailed(uri, status, statusDescription);
}
#endregion
}
class WebRequestFailed : WebRequest {
HttpStatusCode status;
String statusDescription;
Uri itemUri;
public WebRequestFailed(Uri uri, HttpStatusCode status, String statusDescription) {
this.itemUri = uri;
this.status = status;
this.statusDescription = statusDescription;
}
WebException GetException() {
SerializationInfo si = new SerializationInfo(typeof(HttpWebResponse), new System.Runtime.Serialization.FormatterConverter());
StreamingContext sc = new StreamingContext();
WebHeaderCollection headers = new WebHeaderCollection();
si.AddValue("m_HttpResponseHeaders", headers);
si.AddValue("m_Uri", itemUri);
si.AddValue("m_Certificate", null);
si.AddValue("m_Version", HttpVersion.Version11);
si.AddValue("m_StatusCode", status);
si.AddValue("m_ContentLength", 0);
si.AddValue("m_Verb", "GET");
si.AddValue("m_StatusDescription", statusDescription);
si.AddValue("m_MediaType", null);
WebResponseFailed wr = new WebResponseFailed(si, sc);
Exception inner = new Exception(statusDescription);
return new WebException("This request failed", inner, WebExceptionStatus.ProtocolError, wr);
}
public override WebResponse GetResponse() {
throw GetException();
}
public override IAsyncResult BeginGetResponse(AsyncCallback callback, object state) {
Task<WebResponse> f = Task<WebResponse>.Factory.StartNew (
_ =>
{
throw GetException();
},
state
);
if (callback != null) f.ContinueWith((res) => callback(f));
return f;
}
public override WebResponse EndGetResponse(IAsyncResult asyncResult) {
return ((Task<WebResponse>)asyncResult).Result;
}
}
class WebResponseFailed : HttpWebResponse {
public WebResponseFailed(SerializationInfo serializationInfo, StreamingContext streamingContext)
: base(serializationInfo, streamingContext) {
}
}
你必须创建一个HttpWebResponse子类，因为否则无法创建它。
复杂的部分（在GetException()方法中）是如何输入不能重写的值，例如StatusCode。这就是我们最好的伙伴SerializationInfo派上用场的地方！在这里，您可以提供无法重写的值。显然，重写HttpWebResponse的其余部分以使其达到预期效果。
我如何获取所有这些AddValue()调用中的“名称”？从异常消息中！异常消息很友好地按顺序告诉了我每个名称，直到我满意为止。
现在，编译器会抱怨“已过时”，但这仍然可以在.NET Framework版本4中使用。以下是一个（通过）的参考测试用例：
[TestMethod, ExpectedException(typeof(WebException))]
public void WebRequestFailedThrowsWebException() {
string TestURIProtocol = TestContext.TestName;
var ResourcesBaseURL = TestURIProtocol + "://resources/";
var ContainerBaseURL = ResourcesBaseURL + "container" + "/";
WebRequest.RegisterPrefix(TestURIProtocol, new WebRequestFailedCreate(HttpStatusCode.InternalServerError, "This request failed on purpose."));
WebRequest wr = WebRequest.Create(ContainerBaseURL);
try {
WebResponse wrsp = wr.GetResponse();
using (wrsp) {
Assert.Fail("WebRequest.GetResponse() Should not have succeeded.");
}
}
catch (WebException we) {
Assert.IsInstanceOfType(we.Response, typeof(HttpWebResponse));
Assert.AreEqual(HttpStatusCode.InternalServerError, (we.Response as HttpWebResponse).StatusCode, "Status Code failed");
throw we;
}
}
[TestMethod]
[HostType("Moles")]
[Description("Tests that the default scraper returns the correct result")]
public void Scrape_KnownUrl_ReturnsExpectedValue()
{
var mockedWebResponse = new MHttpWebResponse();
MHttpWebRequest.AllInstances.GetResponse = (x) =>
{
return mockedWebResponse;
};
mockedWebResponse.StatusCodeGet = () => { return HttpStatusCode.OK; };
mockedWebResponse.ResponseUriGet = () => { return new Uri("http://www.google.co.uk/someRedirect.aspx"); };
mockedWebResponse.ContentTypeGet = () => { return "testHttpResponse"; };
var mockedResponse = "<html> \r\n" +
" <head></head> \r\n" +
" <body> \r\n" +
" <h1>Hello World</h1> \r\n" +
" </body> \r\n" +
"</html>";
var s = new MemoryStream();
var sw = new StreamWriter(s);
sw.Write(mockedResponse);
sw.Flush();
s.Seek(0, SeekOrigin.Begin);
mockedWebResponse.GetResponseStream = () => s;
var scraper = new DefaultScraper();
var retVal = scraper.Scrape("http://www.google.co.uk");
Assert.AreEqual(mockedResponse, retVal.Content, "Should have returned the test html response");
Assert.AreEqual("http://www.google.co.uk/someRedirect.aspx", retVal.FinalUrl, "The finalUrl does not correctly represent the redirection that took place.");
}
var httpWebResponse = Substitute.For<HttpWebResponse>();
httpWebResponse.StatusCode.Returns(HttpStatusCode.NotFound);
httpWebResponse.StatusDescription.Returns("Not Found");
这不是一个完美的解决方案，但之前对我有效，并且由于其简单性而值得额外关注：
还有一个在Typemock论坛中记录的示例：
using System;
using System.IO;
using System.Net;
using NUnit.Framework;
using TypeMock;
namespace MockHttpWebRequest
{
public class LibraryClass
{
public string GetGoogleHomePage()
{
HttpWebRequest request = (HttpWebRequest)WebRequest.Create("http://www.google.com");
HttpWebResponse response = (HttpWebResponse)request.GetResponse();
using (StreamReader reader = new StreamReader(response.GetResponseStream()))
{
return reader.ReadToEnd();
}
}
}
[TestFixture]
[VerifyMocks]
public class UnitTests
{
private Stream responseStream = null;
private const string ExpectedResponseContent = "Content from mocked response.";
[SetUp]
public void SetUp()
{
System.Text.UTF8Encoding encoding = new System.Text.UTF8Encoding();
byte[] contentAsBytes = encoding.GetBytes(ExpectedResponseContent);
this.responseStream = new MemoryStream();
this.responseStream.Write(contentAsBytes, 0, contentAsBytes.Length);
this.responseStream.Position = 0;
}
[TearDown]
public void TearDown()
{
if (responseStream != null)
{
responseStream.Dispose();
responseStream = null;
}
}
[Test(Description = "Mocks a web request using natural mocks.")]
public void NaturalMocks()
{
HttpWebRequest mockRequest = RecorderManager.CreateMockedObject<HttpWebRequest>(Constructor.Mocked);
HttpWebResponse mockResponse = RecorderManager.CreateMockedObject<HttpWebResponse>(Constructor.Mocked);
using (RecordExpectations recorder = RecorderManager.StartRecording())
{
WebRequest.Create("http://www.google.com");
recorder.CheckArguments();
recorder.Return(mockRequest);
mockRequest.GetResponse();
recorder.Return(mockResponse);
mockResponse.GetResponseStream();
recorder.Return(this.responseStream);
}
LibraryClass testObject = new LibraryClass();
string result = testObject.GetGoogleHomePage();
Assert.AreEqual(ExpectedResponseContent, result);
}
[Test(Description = "Mocks a web request using reflective mocks.")]
public void ReflectiveMocks()
{
Mock<HttpWebRequest> mockRequest = MockManager.Mock<HttpWebRequest>(Constructor.Mocked);
MockObject<HttpWebResponse> mockResponse = MockManager.MockObject<HttpWebResponse>(Constructor.Mocked);
mockResponse.ExpectAndReturn("GetResponseStream", this.responseStream);
mockRequest.ExpectAndReturn("GetResponse", mockResponse.Object);
LibraryClass testObject = new LibraryClass();
string result = testObject.GetGoogleHomePage();
Assert.AreEqual(ExpectedResponseContent, result);
}
}
}